Spotify Top 100 Songs of 2010-2019

import numpy as np import pandas as pd import seaborn as sns import matplotlib.pyplot as plt from wordcloud import WordCloud import re df = pd.read_csv('/work/datasets/Spotify 2010 - 2019 Top 100.csv')

titleobject

Paradise0.3%

944 others99.4%

Missing0.3%

artistobject

Taylor Swift2.1%

443 others97.6%

Missing0.3%

STARSTRUKK (feat. Katy Perry)

3OH!3

My First Kiss (feat. Ke$ha)

3OH!3

I Need A Dollar

Aloe Blacc

Airplanes (feat. Hayley Williams of Paramore)

B.o.B

Nothin' on You (feat. Bruno Mars)

B.o.B

Magic (feat. Rivers Cuomo)

B.o.B

The Time (Dirty Bit)

Black Eyed Peas

Imma Be

Black Eyed Peas

Talking to the Moon

Bruno Mars

Just the Way You Are

Bruno Mars

print(df.info()) total_nan_values = df.isna().sum() print ("Total Number of NaN values:") print(total_nan_values)

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 1003 entries, 0 to 1002
Data columns (total 17 columns):
 #   Column         Non-Null Count  Dtype  
---  ------         --------------  -----  
 0   title          1000 non-null   object 
 1   artist         1000 non-null   object 
 2   top genre      1000 non-null   object 
 3   year released  1000 non-null   float64
 4   added          1000 non-null   object 
 5   bpm            1000 non-null   float64
 6   nrgy           1000 non-null   float64
 7   dnce           1000 non-null   float64
 8   dB             1000 non-null   float64
 9   live           1000 non-null   float64
 10  val            1000 non-null   float64
 11  dur            1000 non-null   float64
 12  acous          1000 non-null   float64
 13  spch           1000 non-null   float64
 14  pop            1000 non-null   float64
 15  top year       1000 non-null   float64
 16  artist type    1000 non-null   object 
dtypes: float64(12), object(5)
memory usage: 133.3+ KB
None
Total Number of NaN values:
title            3
artist           3
top genre        3
year released    3
added            3
bpm              3
nrgy             3
dnce             3
dB               3
live             3
val              3
dur              3
acous            3
spch             3
pop              3
top year         3
artist type      3
dtype: int64

df = df.dropna(axis=0) df

titleobject

Paradise0.3%

Don't0.3%

943 others99.4%

artistobject

Taylor Swift2.1%

Drake1.8%

442 others96.1%

STARSTRUKK (feat. Katy Perry)

3OH!3

My First Kiss (feat. Ke$ha)

3OH!3

I Need A Dollar

Aloe Blacc

Airplanes (feat. Hayley Williams of Paramore)

B.o.B

Nothin' on You (feat. Bruno Mars)

B.o.B

Magic (feat. Rivers Cuomo)

B.o.B

The Time (Dirty Bit)

Black Eyed Peas

Imma Be

Black Eyed Peas

Talking to the Moon

Bruno Mars

Just the Way You Are

Bruno Mars

df.groupby('top genre')['top genre'].count().sort_values(ascending=False)

genres_year = df[['top genre', 'top year']] genres_year = genres_year.rename(columns={'top year':'top_year'}) def cont_gen(ds): years = range(2010, 2020) y_g = pd.DataFrame() for i in years: col = ds[ds.top_year == i]['top genre'].value_counts() col = col[:3] col.to_frame() col = col.rename(str(i)) y_g = y_g.append(col) return y_g genres_years = cont_gen(genres_year) genres_years

atl hip hopfloat64

7.0 - 11.0

barbadian popfloat64

5.0 - 5.0

2010

2011

nan

2012

nan

2013

nan

2014

nan

2015

nan

2016

nan

2017

nan

2018

nan

2019

nan

f1 = plt.figure() colors = ['tomato','lightseagreen', 'Aqua', 'Firebrick', 'DarkViolet', 'Gainsboro', 'Olive', 'Gold', 'Pink', 'Black', 'Blue', 'Red'] plt.style.use('seaborn-darkgrid') genres_years.plot(kind='bar', ax=f1.gca(), stacked=True ,figsize=(10,9), color=colors, ylabel='Number of songs in the top', xlabel='Years', title='Most popular genres of the decade') plt.legend(loc='center left', bbox_to_anchor=(1.0, 0.5)) plt.show()

artist = df.query('`year released` < 2010') artist

titleobject

STARSTRUKK (feat. Katy Perry)4%

Imma Be4%

23 others92%

artistobject

Lady Gaga12%

David Guetta8%

16 others80%

If We Ever Meet Again (Featuring Katy Perry)

Timbaland

Say Aah (feat. Fabolous)

Trey Songz

BedRock

Young Money

178

Good Life

OneRepublic

982

Bohemian Rhapsody - Remastered 2011

Queen

artist.query('`top year` != 2010')

titleobject

artistobject

178

Good Life

OneRepublic

982

Bohemian Rhapsody - Remastered 2011

Queen

df.groupby('artist')['artist'].count().sort_values(ascending=False)

artist_decade = df[['artist', 'top year']] artist_decade = artist_decade.rename(columns={'top year':'top_year'}) def cont_art(ds): years = range(2010, 2020) a_d = pd.DataFrame() for i in years: col = ds[ds.top_year == i]['artist'].value_counts() col = col[:3] col.to_frame() col = col.rename(str(i)) a_d = a_d.append(col) return a_d artist_decades = cont_art(artist_decade) artist_decades

Jason Derulofloat64

3.0 - 3.0

Keshafloat64

4.0 - 4.0

2010

2011

nan

2012

nan

2013

nan

2014

nan

2015

nan

2016

nan

2017

nan

2018

nan

2019

nan

f2 = plt.figure() colors = ['tomato','lightseagreen', 'Aqua', 'Firebrick', 'DarkViolet', 'Gainsboro', 'Olive', 'Gold', 'Pink', 'Black', 'Blue', 'Red', 'Goldenrod', 'mistyrose', 'tan', 'azure', 'gray', 'yellow', 'white', 'greenyellow', 'rosybrown', 'royalblue', 'bisque'] plt.style.use('seaborn-darkgrid') artist_decades.plot(kind='bar', ax=f2.gca(), stacked=True ,figsize=(10,9), color=colors, title='Most popular artists of the decade', ylabel='Number of songs in the top', xlabel='Years') plt.legend(loc='center left', bbox_to_anchor=(1.0, 0.5)) plt.show()

sad = df[['val', 'top year']] sad = sad.rename(columns={'top year':'top_year'}) def cont_sad(ds): years = range(2010, 2020) s = {} for i in years: sa = sad[sad.top_year == i] pro = sa.mean() s[i] = pro[0] return s sad1 = cont_sad(sad) sad1 = pd.DataFrame(sad1.items(), columns=['year', 'prom_sad']) sad1

yearint64

2010 - 2019

prom_sadfloat64

46.82 - 56.74

2010

56.74

2011

54.85

2012

53.29

2013

53.05

2014

50.95

2015

46.82

2016

46.92

2017

47.72

2018

47.56

2019

51.11

sns.set_theme(style="darkgrid") f3 = sns.lineplot(data=sad1, x="year", y="prom_sad") f3.set_title('Graph of the change in the mood of the songs in the decade 2010-2019') f3.set_ylabel('Positivity of the songs') f3.set_xlabel('Years') plt.show()

nrgy = df[['nrgy', 'top year']] nrgy = nrgy.rename(columns={'top year':'top_year'}) def cont_nrgy(ds): years = range(2010, 2020) s = {} for i in years: sa = nrgy[nrgy.top_year == i] pro = sa.mean() s[i] = pro[0] return s energy = cont_nrgy(nrgy) energy = pd.DataFrame(energy.items(), columns=['year', 'prom_energy']) energy

yearint64

2010 - 2019

prom_energyfloat64

62.79 - 76.65

2010

76.65

2011

76.1

2012

74.07

2013

72.44

2014

71.03

2015

67.53

2016

65.8

2017

63.16

2018

65.45

2019

62.79

sns.set_theme(style="darkgrid") f4 = sns.lineplot(data=energy, x="year", y="prom_energy") f4.set_title('Graph of the change of energy of the songs in the decade 2010-2019') f4.set_ylabel('Energy of the songs') f4.set_xlabel('Years') plt.show()

dnce = df[['dnce', 'top year']] dnce = dnce.rename(columns={'top year':'top_year'}) def cont_dnce(ds): years = range(2010, 2020) s = {} for i in years: sa = dnce[dnce.top_year == i] pro = sa.mean() s[i] = pro[0] return s dance = cont_dnce(dnce) dance = pd.DataFrame(dance.items(), columns=['year', 'prom_dance']) dance

yearint64

2010 - 2019

prom_dancefloat64

63.55 - 73.58

2010

65.29

2011

63.56

2012

64.01

2013

63.55

2014

65.88

2015

65.22

2016

64.53

2017

73.58

2018

70.8

2019

72.34

sns.set_theme(style="darkgrid") f5 = sns.lineplot(data=dance, x="year", y="prom_dance") f5.set_title('Graph of the change of danceability of the songs in the decade 2010-2019') f5.set_ylabel('Danceability of the songs') f5.set_xlabel('Years') plt.show()

text = " ".join(title for title in df.title) text = "".join(re.split("$|$|\[|\]", text)[::2]) text = text.replace('Edit ', '') text = text.replace('Radio ', '') text = text.replace('Remix', '') word_cloud = WordCloud(collocations = False, background_color = 'white').generate(text) plt.imshow(word_cloud, interpolation='bilinear') plt.title('Map of the most common words in song titles') plt.axis("off") plt.show()

sns.set_theme(style="darkgrid") f5 = sns.lineplot(data=dance, x="year", y="prom_dance") f4 = sns.lineplot(data=energy, x="year", y="prom_energy") f3 = sns.lineplot(data=sad1, x="year", y="prom_sad") f5.set_xlabel('Years') plt.show()

df1 = df.rename(columns={'top year':'top_year'}) df2 = df1[df1.top_year > 2016] df2

titleobject

Bodak Yellow0.7%

Him & I (with Halsey)0.7%

292 others98.7%

artistobject

Post Malone3.3%

Drake2.7%

176 others94%

770

Slow Hands

Niall Horan

771

Magnolia

Playboi Carti

772

Chameleon

PNAU

773

Feel It Still

Portugal. The Man

774

Congratulations

Post Malone

775

I Fall Apart

Post Malone

776

Swang

Rae Sremmurd

777

Black Beatles

Rae Sremmurd

778

Chill Bill (feat. J. Davi$ & Spooks)

Rob $tone

779

Losin Control

Russ

df2 = df2[df2.dnce > 66] df2 = df2.reset_index()

df2['dnce'].sort_values(ascending=False)

df2.iloc[[106, 153, 37, 5, 46]]

indexint64

titleobject

106

856

Yes Indeed

153

923

Money

750

Gucci Gang

705

Caroline

764

Bad and Boujee (feat. Lil Uzi Vert)

expr = df[['dnce', 'nrgy', 'val']] sns.heatmap(expr.corr(), annot = True)